Vol. 14: 127–134, 2011 ENDANGERED SPECIES RESEARCH Published online June 22 doi: 10.3354/esr00350 Endang Species Res

Application of the precautionary principle to taxa of uncertain status: the case of the Bellinger River turtle

Arthur Georges1,*, Ricky-John Spencer1, 2, Michael Welsh3, H. Bradley Shaffer4, Rachael Walsh1, Xiuwen Zhang1

1Institute for Applied Ecology, University of Canberra, ACT 2601, 2Native and Pest Animal Unit, School of Natural Sciences, University of Western Sydney, Locked Bag 1797, Penrith South, 2750, Australia 3Michael Welsh Fauna Consultant, 7 Chisholm Road, Gerringong, New South Wales 2534, Australia 4Department of Evolution and Ecology, University of California, One Shields Ave, Davis, California 95616, USA

ABSTRACT: Legislative and regulatory frameworks for conserving biodiversity often focus on the species as a fundamental unit for protection. In cases where the taxonomic or conservation status of a species is uncertain, the precautionary principle may be invoked in listing suspected but as yet undescribed taxa as vulnerable or endangered. In this paper, we present an evaluation of the taxo- nomic status of what has been regarded as a distinctive but as yet undescribed species of freshwater turtle, declared endangered in the Action Plan for Australian Reptiles and vulnerable in the sched- ules of state and federal conservation acts. Using mitochondrial sequence variation, we show that the Bellinger River turtle is an unremarkable population of a common and widespread species, Emydura macquarii. In addition, we present evidence suggesting that it may have been recently introduced to, or may be a recent invader of, the Bellinger River (New South Wales, Australia) where it may come to compete with Myuchelys georgesi, an endemic found only in the Bellinger River. Our study illus- trates the need to couple fundamental research with on-ground action early in an adaptive manage- ment context, particularly where taxonomic status of the target species is uncertain. Short-term cost savings of failing to do so may come to be greatly exceeded by longer-term opportunity loss where conservation dollars are limited.

KEY WORDS: Taxonomic impediment · Emydura macquarii · Myuchelys georgesi · Elseya · Mitochondrial sequence variation

Resale or republication not permitted without written consent of the publisher

INTRODUCTION Biodiversity encapsulates diversity of biotic commu- nities, diversity of species in those communities, Biodiversity loss is one of the greatest challenges of genetic diversity within species, and the ecological and the 21st century (McGlothlin 2004). Changes brought evolutionary processes that create and maintain that about by growth of human populations and per capita diversity (Wilson & Peter 1988, World Resources Insti- consumption are driving many changes, with adverse tute 1992, de Long 1996). However, in many legislative consequences for wildlife on all continents, through and regulatory systems, the species remains the pri- the combined influences of over-exploitation (Lauck et mary focus for conservation (Hey et al. 2003). As new al. 1998, Rosenberg 2003), deforestation, habitat species are discovered, rapid assessment of their status degradation, fragmentation or loss (Fahrig 1997, becomes a priority so that measures can be taken to Yamaura et al. 2011), exotic predators and diseases ensure that they are protected (Pauly et al. 2007). In (Berger et al. 1998, Daszak et al. 1999, Lips et al. 2006) some jurisdictions, such as Australia, the precautionary and climate change (Walther et al. 2002). principle (Raffensperger et al. 1999) is used to allow

*Email: [email protected] © Inter-Research 2011 · www.int-res.com 128 Endang Species Res 14: 127–134, 2011

protection and remediation to be initiated in advance Bellinger River Emydura and use mitochondrial DNA of formal description of ‘species’ thought to be at risk. sequences of Emydura spp. and the endemic Bellinger The Australian Environmental Protection and Biodi- River sawshell turtle Myuchelys georgesi to address versity Conservation (EPBC) Act and most state juris- these 2 questions. Taxonomic research is in decline dictions allow nomination of populations regarded as, despite the biodiversity imperative (Hopkins & Freck- or suspected to be, distinct species or subspecies, some leton 2002), and in Australia is compounded by the of which await formal description. This practice leads proliferation of names attached to taxonomic concepts to satisfactory conservation outcomes when the need with little or no support (see Georges & Thomson 2010, for management intervention is pressing, but the infor- Oliver & Lee 2010) and uncritically adopted by state mation needed to make sound taxonomic judgements and federal agencies. The role of a well-founded tax- has yet to be fully realised. The risk is that conserva- onomy in underpinning conservation is well estab- tion resources may be allocated on taxa that later turn lished (Mace 2004). Our study serves to highlight the out to be of relatively low conservation value (Zhu et importance of a commitment to fundamental taxo- al. 1998). Such would occur when the initial judgement nomic research, concurrent with the early stages of on the validity of a taxon is not borne out by sub- precautionary management intervention, to address sequent investigation. the uncertainty in knowledge that leads to the invoca- Here we present an example where the precaution- tion of the precautionary principle in the listing of new ary principle was applied to what was, on the basis of taxa whose taxonomic status is uncertain. anecdotal evidence, a ‘species’ at risk — a distinctive form of freshwater turtle found only in the Bellinger River of eastern Australia (Cann 1998). Listed as vul- MATERIALS AND METHODS nerable in the Reptile Action Plan (Cogger et al. 1993) as Emydura signata, vulnerable in the Schedules of the Turtles were collected between 1 and 17 April 2007 Threatened Species Act, New South Wales (NSW in the Bellinger and Kalang Rivers (New South Wales, NPWS 2001) as E. macquarii, Bellinger River Form, Australia) (Fig. 1) by hand with mask and fins, in and vulnerable in the national EPBC Act (Common- cathedral traps (Legler 1960, as modified by Georges wealth of Australia 1999) as E. macquarii signata et al. 2006) baited with liver and sardines, and with (Bellinger River, NSW), this ‘species’ of turtle was used dipnets by spotlight at night. Traps were checked at as a flagship to marshal community support for a range intervals of 4 to 10 h, re-baited after approximately of initiatives in riparian and riverine restoration and 24 h, and left in place for 24 to 48 h at each location. invasive fox control (NSW NPWS 2001). Discovered in Mature males were distinguished from mature fe- 1990 in a single waterhole in the middle reaches of the males based on tail length (Georges et al. 2006), and the Bellinger River by P. King (King & Heatwole 1994) as minimum size of visually identifiable males (118 mm part of broader studies and again at the same location carapace length) was used as the upper size limit for all in 1992 by Cann (1998), captures of the Bellinger Emy- juveniles. A small sliver of tissue was taken from the out- dura were initially ex tremely rare. Recent surveys side edge of the vestigial toe of the rear foot, stored in funded by the NSW Government have shown progres- 75% ethanol for DNA analysis, and later transferred to sive increases in abundance and distribution in the –20°C storage freezers. Each turtle was photographed Bellinger River, to the point where the turtles are now (dorsal and ventral view, lateral view of the head) and widespread and common in the drainage. However, presence or absence of the cervical scute and/or head the question of whether the Bellinger River Emydura is shield was noted, characters which ty pically differ be- a distinctive taxon or an unremarkable representative tween Emydura macquarii and Myuchelys georgesi. All of the more widespread E. macquarii macquarii (sensu turtles were released within 2 h at their point of capture. Georges & Thomson 2010) remains unanswered. So DNA was extracted from the 76 specimens of Emy- too does the more difficult question of whether it is a dura macquarii, 16 Myuchelys georgesi, 3 suspected relict endemic having barely endured and recovered hybrids and 1 suspected specimen of M. latisternum from a catastrophic decline, or whether it is a recent from the Bellinger River drainage (including the introduction to the river, possibly with human assis- Kalang tributary). An additional 34 specimens of E. tance, that has established and is in the early stages of macquarii from 3 localities in the vicinity of Coffs Har- invasion. Both questions are critical to deciding what bour were also extracted as a second phase in the management intervention is required, if any, and the study. Genomic DNA was extracted from skin tissues value of the Bellinger River Emydura as a flagship using salt extraction or Chelex (Bio-Rad) beads as species of conservation concern. described by FitzSimmons et al. (1997). Sub-samples of In this paper, we report the results of the most recent skin were digested overnight at 55°C in 300 µl of buffer survey into the distribution and abundance of the (40 mM Tris, 20 mmol EDTA, 100 mmol NaCl), 20 µl of Georges et al.: Precautionary principle and undescribed taxa 129

Fig. 1. Bellinger catchment (at the junction of the crosshairs in the inset) showing the capture localities of the Bellinger Emydura (d). The Kalang River is essentially a tributary of the Bellinger River, and is shown on the map as running through Urunga and Brierfield (lighter shading)

10% sodium dodecyl sulphate (SDS) and 10 µl of 10 mg and 4 bp (Table 1). In contrast, the endemic Myuchelys ml–1 Proteinase K. DNA was salt extracted and precip- georgesi yielded only a single haplotype (Genbank itated with 150 µl of 7.5 mol ammonium acetate and 1 accession nos. JN088723–30). The sequences gener- ml of 100% ethanol, washed with 70% ethanol, and ated by the UC Davis DNA sequencing facility yielded then re-suspended in 50 µl of water. A 4 µl subsample a further 68 unique haplotypes from across the range of each extraction was run out on a gel to confirm the of the southern Emydura (Table 2). One of these haplo- success of the extraction. PCR used the control region types was for a Bellinger Emydura collected by P. King, (CR) primers TCR500 and EMYTHR (Engstrom et al. and its haplotype was identical to another Bellinger 2004). PCR products (50 µl of each sample) were Emydura from the present study. precipitated with 50 µl of 20% polyethylene glycol The haplotypes of the Bellinger River fell within a (PEG), washed with 80% ethanol and re-suspended clade comprising the Bellinger, Clarence, Macleay, in 13 µl of water. The cleanup PCR products were Nambucca and Hastings drainages. None of the 6 hap- packaged and sent to Macrogen (Seoul) for sequenc- lotypes identified in the Bellinger were found in ing. The resulting sequences were checked, analysed drainages outside this small collection of adjacent and aligned using Geneious Pro 4.6.4. Additional CR drainages (Table 2) sequences for a range of populations of the southern The network diagram generated for the Bellinger Emydura were sequenced at the UC Davis DNA Emydura haplotypes shows the most parsimonious sequencing facility by H. B. S. using standard protocols evolutionary relationship among these haplotypes, (e.g. Spinks et al. 2010). with 4 of the 6 haplotypes common, and haplotype HP9 Unique mitochondrial haplotypes were identified as the most parsimonious choice of the ancestral haplo- using FaBox (Villesen 2007). A phylogeny for those haplotypes was constructed under the maximum parsi- Table 1. Emydura macquarii. Uncorrected distances between mony criterion and standard default settings using the control region haplotypes (HP) found in turtles from the computer package PAUP* (Swofford 1996) and 2 Bellinger River and adjacent drainages. The lower matrix shows raw base pair differences; the upper matrix shows per- Northern Territory Emydura victoriae haplotypes as centage differences (N = 428 bp). Haplotype 10 is a related outgroups. The full analysis will be presented else- haplotype from the Clarence River that was not detected in where (H. B. Shaffer et al. unpublished). A haplotype the Bellinger River network was constructed for the Bellinger River haplo- types using the programme TCS (Clement et al. 2000). HP7 HP8 HP9 HP10 HP11 HP12 HP13

HP7 0.24 0.47 0.71 0.71 0.94 0.71 HP8 1 0.24 0.47 0.47 0.71 0.48 RESULTS HP9 2 1 0.24 0.24 0.47 0.24 HP10 3 2 1 0.47 0.71 0.47 Analysis of the 428 bp CR haplotypes of the 76 spec- HP11 3 2 1 2 0.24 0.48 HP1243231 0.72 imens of Emydura macquarii from the Bellinger River HP13321223 yielded 6 unique haplotypes differing by between 1 130 Endang Species Res 14: 127–134, 2011

Table 2. Emydura macquarii. Haplotypes (HP) ordered by locality for 428 bp of mitochondrial control region for the southern Emydura (cf macquarii). Bellinger River haplotypes are in bold. MDB: Murray-Darling Basin

Drainage Haplotypes Drainage Haplotypes

New South Wales Bellinger R. HP7, HP8, HP9, HP11, HP12, HP13 MDB Murray R. HP43, HP44, HP45, HP46 Bulloo R. HP47, HP73 MDB Murrumbidgee R. HP42, HP44 Clarence R. HP9, HP10, HP13, HP52 MDB Namoi R. HP48 Hastings R. HP8 Nambucca R. HP8 Hunter R. HP1, HP2, HP3, HP4 Paroo R. HP51 Macleay R. HP8 Richmond R. HP15 MDB Border R. HP46, HP47, HP48, HP54 Richmond R. HP16 MDB Lachlan R. HP46 Richmond R. HP17 MDB Macquarie R. HP50 Richmond R. HP18 Queensland Brisbane R. HP33, HP52 Johnstone R. HP39, HP40 Bulloo R. HP72 Herbert R. HP22 Burdekin R. HP26, HP27, HP28, HP29, Johnstone R. HP20 HP30, HP31, HP37 Burnett R. HP55, HP58, HP59, HP61 Mary R. HP49, HP59, HP60, HP62, HP64, HP65, HP67, HP68 Cooper Creek HP70, HP71, HP72, HP73 MDB Condamine-Balonne HP48 Endeavour R. HP25 MDB Warrego R. HP41, HP47, HP74 Fitzroy-Dawson R. HP14, HP34, HP35, HP36, HP38, HP53, Proserpine HP37 HP56, HP57, HP63, HP66, HP69 Fraser Is. HP32 Ross R. HP19, HP20, HP23, HP24 Hahn R. HP20, HP21 Tweed R. HP64

type (Fig. 2). Of the Bellinger haplotypes, HP13 and tween the Bellinger River proper and the Kalang River, HP9 are shared with the Clarence River, HP8 (and pos- complete riverine connectivity within the Bellinger sibly the 1 bp variant HP7) are shared with the River and independent ecological evidence of substan- Macleay, Nambucca and Hastings Rivers. Haplotypes tial turtle movements within the system (Spencer 2006) HP11 and HP12 were unique to the Bellinger drainage. would argue against spatial genetic differentiation The distribution of haplotypes within the Bellinger within the Bellinger River. Yet we are seeing such dif- River drainage basin is also instructive (Fig. 3). While ferentiation. Of particular interest are the 2 haplotypes one might expect some genetic differentiation be - (HP11 and HP12) that were not detected in our sam- ples from outside the Bellinger drainage. All 15 speci- mens bearing these haplotypes were found in a single waterhole 750 m upstream from the centre of the town- ship of Bellingen. In a small, continuously flowing,

Fig. 2. Emydura macquarii. Maximum parsimony network showing the relationship among haplotypes from the Bel - linger River. Area of the circles is proportional to the fre- quency of the haplotypes in the population (n = 72 turtles). Small black dots represent transitionary haplotypes not rep- Fig. 3. Emydura macquarii. Distribution of haplotypes within resented in our sample. Network constructed using the the Bellinger River drainage (n = 72). Vertical axis shows per- computer programme ‘TCS’ (Clement et al. 2000) centage representation for each haplotype across localities Georges et al.: Precautionary principle and undescribed taxa 131

highly connected river system, the most plausible Blamires et al. 2005), but this resulted in only 11 new in- explanation for this is a recent introduction of more dividuals captured, all upstream of Thora and close to than 1 specimen. In 2000, 3 of several Emydura Ralph’s Crossing. The survey we report in this paper, brought to the Bellingen township from a seizure in conducted in 2007, captured 76 individuals distributed escaped into the river (B. Nesbitt pers. throughout the drainage, a substantial proportion of comm.). It may be that these are the source of these which were less than 10 yr old (Spencer et al. 2007). localised haplotypes, which have not yet had time to Notwithstanding variation in the methods and efficien- propagate throughout the system. Subsequent screen- cies of these surveys, there is little doubt that the ing of populations in the vicinity of the township of Bellinger Emydura has moved from a situation of criti- Coffs Harbour in search of the source of these speci- cally low numbers to a more secure population from the mens yielded 3 haplotypes (15 individuals, 3 localities, time of first discovery to the present. The critical issue is HP9, HP12 and HP13). Note that HP12 was found in whether these population trends are of a distinctive en- the Bellinger River only at the release site near Bellin- demic population recovering from the brink of extinc- gen township. tion, or the consolidation and rapid population growth of a regionally abundant taxon recently introduced to the Bellinger River. From a conservation perspective, is DISCUSSION there cause to rejoice or cause for concern? Our genetic data do not support the proposition that River turtles of the genus Emydura are widely dis- the Bellinger Emydura is a distinctive species or sub- tributed in Australia and southern New Guinea, wher- species (Cann 1998), nor do they support its listing as a ever permanent waters are found, with the exception threatened taxon in the Action Plan for Australian Rep- of the south-western corner of the mainland Australia tiles, the Schedules of the Threatened Species Act, and Tasmania (Georges & Thomson 2010). Their spe- NSW, or the schedules of the national EPBC Act. cies-level taxonomy has been controversial, and issues Indeed, the Bellinger Emydura is genetically unre- of species delineation remain unresolved in tropical markable, with very little mitochondrial genetic diver- northern Australia. Southeastern and eastern Australia gence between the Bellinger populations of E. mac- appear to be occupied by a single widespread poly- quarii and those of the adjacent Macleay, Hastings, typic species (Georges & Adams 1996) that has vari- Nambucca and Clarence drainages (Table 1). There is ously been divided into species and subspecies (Cann no evidence in our data to suggest that the Bellinger 1998, McCord et al. 2003). One recent classification form should be regarded as a separate biological (Georges & Thomson 2010) has E. macquarii macquarii species, a sub-species, or indeed a distinct evolutionar- as occupying the Murray-Darling drainage and adja- ily significant unit (ESU; sensu Moritz 1994). The cent coastal drainages from the Brisbane River south to Bellinger population shares haplotypes with adjacent the Hawkesbury- of the Sydney region. systems to the north and south and so shares haplo- These east coastal populations are quite divergent types with populations regarded as unremarkable genetically (H.B. Shaffer et al. unpublished), differ in representatives of the widespread southern Emydura minor morphological attributes (Cann 1998), and vary (Georges & Thomson 2010). The diagnostic morpho- dramatically in maximum body size (Cann 1998). They logical features identified by Cann (1998), such as the have been referred to variously as E. signata (Iverson flaring of the carapace and the colour of the iris, are 1992), E. macquarii signata (Cogger 2000) or as a series both highly variable among populations of the south- of single-drainage subspecies (Cann 1998). The ern Emydura, and no rigorous morphological analysis Bellinger Emydura has been regarded as one of these has been published to support the diagnosis. There is distinctive forms of Emydura signata or E. macquarii. no evidence to suggest that the Bellinger population of There is evidence that the Bellinger Emydura has re - E. macquarii is anything more than a population of a cently increased in abundance and distribution within widespread species occupying sub-optimal habitat the Bellinger drainage. In the 15 to 20 yr prior to the (simple channel, oligotrophic, clear-water river flow- publication of the Action Plan for Australian Reptiles ing over a gravel bed, with heavily forested banks). (Cogger et al. 1993), opportunistic widespread search- The proposition that the Bellinger Emydura is an iso- ing for this taxon along the length of the Bellinger River lated relict endemic species or subspecies (Cogger et (King & Heatwole 1994, Cann 1998, Allanson & al. 1993, Cann 1998), now in the early stages of recov- Georges 1999) had its distribution restricted to 1 local- ery, is without support, based on the limited genetic ity, Ralph’s Crossing, 11 km upstream of Thora. Be- data presented here. tween 2000 and 2005, the river was both opportunisti- Our data are consistent with the proposition that some, cally and systematically searched annually for if not all, of the genetic variation within the Bellinger Bellinger River Emydura (Spencer & Thompson 2000, Emydura is a result of recent introductions or invasions. 132 Endang Species Res 14: 127–134, 2011

Distinguishing between human-induced introduction or distinctive Myuchelys georgesi, which is endemic to natural invasion from adjacent drainages is more prob- this small coastal drainage (Georges 1993), the conser- lematic because any evidence is by necessity inferential vation initiatives linked to the Bellinger River Emydura and the 2 possibilities are not mutually exclusive. On the are of immense value in achieving conservation objec- basis of the haplotype analysis, we can confidently elim- tives in the catchment. Downplaying the status of the inate the commercial pet trade operating legally in South flagship species presents a particular problem for man- Australia (JB’s Turtles, Barmera) as a possible source of agement. We recommend that the emphasis on the introductions, and indeed can confidently eliminate any Bellinger Emydura be broadened to also encompass the of the drainages occupied by the southern Emydura out- endemic M. georgesi. This shift in focus should be easy side those in the neighbouring region of the Bellinger to achieve without putting at risk valuable community (Clarence, Macleay, Nambucca, Hastings). It is unusual support for fox control and restoration of riparian vege- to see such diversity in mitochondrial haplotypes in a tation. Subsequently, the emphasis can be further species occupying a drainage as small as the Bellinger (6 shifted to the local endemic M. georgesi, should this be mitochondrial haplotypes, 4 of which are well repre- required as more definitive information becomes avail- sented), and in a population as small as that suspected able, particularly with regard to the risks associated for the Bellinger Emydura. Contrast this with the single with competition and potential hybridisation between haplotype for the much more abundant Myuchelys geor- E. macquarii and M. georgesi. gesi in the same drainage. Three of the 4 common Emy- There is a salient lesson in the processes leading up dura haplotypes are shared with adjacent drainages. to the declaration of the Bellinger Emydura as a listed Given these observations, our interpretation is that the taxon of conservation concern in state and federal reg- most likely scenario is that some of the Bellinger haplo- ulations. Our view is not that a mistake was made in types are derived from the more substantial turtle popu- drawing upon the precautionary principle at the time lations in adjacent drainages, either through natural dis- of listing the Bellinger Emydura, which has apparently persal or through human-mediated movement of turtles. ended up protecting a population of low conservation The reliable reports of Emydura being released into the status. The precautionary principle followed by an Bellinger River from adjacent catchments (Coffs Creek adaptive management program (Walters & Holling region; B. Nesbitt pers. comm.), the detection of HP12 in 1990) offers an appropriate way out of the paralysis we Coffs Creek, and the restricted distribution of HP12, might otherwise face when a taxon is suspected to be found only in a single pool adjacent to the Bellingen at grave risk and we understand so little about it that township in a contiguous riverine system, provide we find ourselves unable to identify a truly safe course evidence that at least one such introduction has become of action (Doremus 2001). Such an approach, however, established. should not be a basis for failure to subsequently under- While we cannot conclusively eliminate the possibil- take the fundamental research necessary to reduce ity of natural dispersal events from adjacent drainages, uncertainty in the chosen course of action. In this case, we believe that the most likely scenario is that the the uncertainty arose because of the poor taxonomic Emydura have been introduced to the Bellinger knowledge of the Australian Chelidae. Taxonomic drainage through a series of independent releases; first research is arguably in decline (Hopkins & Freckleton in the vicinity of Ralph’s Crossing in the early 1990s, 2002), in part because of the discipline’s failure to man- and the most recent near the township of Bellingen. age the less productive nomenclatural activity eroding Emydura species favour complex river systems with a perceptions of the value of taxonomic work in the combination of lentic and lotic conditions, including broader scientific community (Godfray 2002, Georges permanent and semi-permanent off-channel bill- & Thomson 2010), but also because governments are abongs, swamps and wetlands. They do less well in less willing to support the scientific functions of their systems with simple river channels and a well-devel- museums and to continue to fund staff with taxonomic oped closed canopy in the riparian zones and adjacent research capacity. Yet with the looming biodiversity forest (Georges et al. 2008). It may be that in either crisis in which turtles appear to be central players (van case, widespread clearing and altered conditions Dijk et al. 2000, Rhodin et al. 2010), a stable nomencla- within the catchment since European settlement have ture for taxonomic concepts that are part of the body of made the river more conducive to the establishment of science could not be more important (Spinks et al. Emydura species. 2009, Georges & Thomson 2010). Without it, resources The Bellinger Emydura has been used as a flagship can be diverted inappropriately (see Smith et al. 1993), ‘species’ upon which a range of conservation initiatives regulations that govern wildlife trade circumvented are based (NSW NPWS 2001). While direct conserva- (Kuchling et al. 2007), biodiversity assessments dis- tion efforts have been arguably misplaced, and perhaps torted (Agapow et al. 2004, Isaac et al. 2004) and con- could have been better directed at the phylogenetically servation effort wasted. The failure in the case of the Georges et al.: Precautionary principle and undescribed taxa 133

Bellinger Emydura was that this fundamental taxo- Acknowledgements. We thank the many people who assisted nomic work was not undertaken soon after the precau- us in the field, but especially J. Faulkner, B. Bailey, C. Eisem- tionary principle was used to declare it as a vulnerable berg and M. de la Cueva. We also thank the community for their hospitality and assistance with locating taxon in the Action Plan for Australian Reptiles (Cog- sites and providing access to those sites. B. Nesbitt and his ger et al. 1993) and in advance of its listing on state and team assisted greatly with local knowledge, companionship national conservation schedules. and assistance in the field. This work was funded by the NSW This case is not isolated. In countries such as Aus- National Parks and Wildlife Service and Bellinger Landcare, and was approved by the University of Western Sydney ani- tralia, where elements of the flora and fauna are mal ethics committee (A6291). The manuscript was prepared diverse and poorly studied, the precautionary principle as part of the Science Writers Club (Kioloa, December 2009), commonly comes into play in assessments of conserva- and we are grateful for the many valuable comments received tion status and listing of taxa under the regulatory from participants. schedules of state and federal biodiversity conserva- tion acts. If the costs of delay appear higher than the LITERATURE CITED costs of potentially mistaken action, opinion rather than evidence has been an appropriate foundation for Agapow PM, Bininda OR, Crandall KA, Gittleman JL, Mace action, and weight of opinion was the basis of many of GM, Marshall JC, Purvis A (2004) The impact of species concept on biodiversity studies. Q Rev Biol 79:161–179 the judgements in the Action Plan for Australian Rep- Allanson M, Georges A (1999) Diet of a sibling species pair of tiles. Processes such as these can subsequently lead to freshwater turtles, Elseya purvisi and Elseya georgesi the development of action and recovery plans, the (Testudinata: Chelidae), from eastern Australia. Chelon- implementation of which are fiscally costly and may ian Conserv Biol 3:473–476 result in opportunity loss if resources are directed Berger L, Speare R, Daszak P, Green DE and others (1998) Chytridiomycosis causes amphibian mortality associated away from more pressing conservation issues. In some with population declines in the rain forests of Australia cases, development proposals can be halted or sub- and Central America. Proc Natl Acad Sci USA 95: stantially modified to meet perceived demands of con- 9031–9036 servation and environmental sustainability, only to find Blamires SJ, Spencer RJ, King P, Thompson MB (2005) Popu- lation parameters and life-table analysis of two co-existing later that the basis for the environmental concerns was freshwater turtles: are the Bellinger River turtle popula- poorly founded. This can serve to erode the commit- tions threatened? Wildl Res 32:339–347 ment of well-minded companies and citizens to appro- Cann J (1998) Australian freshwater turtles. Beaumont Pub- priately factor environmental concerns in their devel- lishing, Singapore opment plans (Doremus 1997) and serves no one well. Clement M, Posada D, Crandall K (2000) TCS: a computer program to estimate gene genealogies. Mol Ecol 9: The need for objective taxonomic and population data 1657–1660 in conservation is critical. A review of the use of anec- Cogger H (2000) Reptiles and amphibians of Australia. Reed dotal data for the conservation of 3 high-profile, rare New Holland, Sydney and endangered species led to large errors of omission Cogger H, Cameron E, Sadlier R, Eggler P (1993) The Action Plan for Australian Reptiles. Commonwealth Department and commission, which greatly influenced the alloca- of Environment, Heritage and Water (formerly Australian tion of limited funds and the efficacy of subsequent Nature Conservation Agency), Canberra conservation efforts (McKelvey et al. 2008). Commonwealth of Australia (1999) Environment Protection We advocate the application of the precautionary prin- and Biodiversity Conservation (EPBC) Act 1999. Depart- ciple followed by an adaptive approach to species con- ment of the Attorney-General, Canberra Daszak P, Berger L, Cunningham AA, Hyatt AD, Green DE, servation. There needs to be better balance between Speare R (1999) Emerging infectious diseases and amphib- strategic research and on-ground management action, ian population declines. Emerg Infect Dis 5:735–748 giving greater emphasis to the former, particularly in the de Long DJ (1996) Defining biodiversity. Wildl Soc Bull very early stages of any adaptive management regimen. 24:738–749 Doremus H (1997) Listing decisions under the Endangered When the issue is regarded as urgent, such as when a Species Act: why better science isn’t always better policy. new species is discovered or suspected, and perceived to Wash Univ Law Q 75:1029 be under threat, the tendency is to do the reverse, viz. to Doremus H (2001) Adaptive management, the Endangered fund action on the ground and neglect the fundamental Species Act, and the institutional challenges of ‘new age’ research of strategic value. The short-term cost savings environmental protection. Washburn Law J 41:50–89 Engstrom TN, Shaffer HB, McCord WP (2004) Multiple data of this approach may ultimately be greatly exceeded by sets, high homoplasy, and the phylogeny of softshell tur- the longer-term opportunity losses where conservation tles (Testudines: Trionychidae). Syst Biol 53:693–710 dollars are limited. Our fundamental conclusion is that Fahrig L (1997) Relative effects of habitat loss and fragmenta- taxonomic research should proceed simultaneously with tion on population extinction. J Wildl Manag 61:603–610 FitzSimmons NN, Limpus CJ, Norman JA, Goldizen AR, conservation and management, if not before, at least in Miller JD, Moritz C (1997) Philopatry of male marine tur- cases where sound management hinges on taxonomic tles inferred from mitochondrial DNA markers. Proc Natl decisions. Acad Sci USA 94:8912–8917 134 Endang Species Res 14: 127–134, 2011

Georges A (1993) Setting conservation priorities for Aus- lication of unnamed new species. Taxon 59:1201–1205 tralian freshwater turtles. In: Lunney D, Ayers D (eds) Her- Pauly GB, Piskurek O, Shaffer HB (2007) Phylogeographic petology in Australia — a diverse discipline. Royal Society concordance in the southeastern United States: the flat- of New South Wales, Sydney, p 49–58 woods salamander, Ambystoma cingulatum, as a test case. Georges A, Adams M (1996) Electrophoretic delineation of Mol Ecol 16:415–429 species boundaries within the short-necked chelid turtles Raffensperger C, Tickner J, Jackson W (1999) Protecting pub- of Australia. Zool J Linn Soc 118:241–260 lic health and the environment: implementing the precau- Georges A, Thomson S (2010) Diversity of Australasian fresh- tionary principle. Island Press, Washington, DC water turtles, with an annotated synonymy and keys to Rhodin AGJ, van Dijk PP, Iverson JB, Shaffer HB (2010) Tur- species. Zootaxa 2496:1–37 tles of the world, 2010 update: annotated checklist of tax- Georges A, Guarino F, White M (2006) Sex-ratio bias across onomy, synonymy, distribution, and conservation status. populations of a freshwater turtle (Testudines: Chelidae) Chelonian Res Monogr 5:000.85–000.164, doi:10.3854/ with genotypic sex determination. Wildl Res 33:475–480 crm.5.000.checklist.v3.2010 Georges A, Alacs E, Pauza M, Kinginapi F, Ona A, Eisemberg Rosenberg AA (2003) Managing to the margins: the overex- C (2008) Freshwater turtles of the Kikori Drainage, Papua ploitation of fisheries. Front Ecol Environ 1:102–106 New Guinea, with special reference to the pig-nosed tur- Smith TB, Bruford MW, Wayne RK (1993) The preservation of tle, Carettochelys insculpta. Wildl Res 35:700–711 process: the missing element of conservation programs. Godfray HCJ (2002) Challenges for taxonomy. Nature 417: Biodivers Lett 1:164–167 17–19 Spencer RJ (2006) The Bellinger River Emydura survey: Hey J, Waples RS, Arnold ML, Butlin RK, Harrison RG (2003) effects of foxes and interspecific competition. Report to Understanding and confronting species uncertainty in NSW National Parks and Wildlife Service, Dorrigo biology and conservation. Trends Ecol Evol 18:597–603 Spencer R, Thompson MB (2000) The Bellinger River Emy- Hopkins GW, Freckleton RP (2002) Declines in the numbers of dura survey: the ecology and status of Emydura macquarii amateur and professional taxonomists: implications for and Elseya georgesi in the Bellinger River. Report to NSW conservation. Anim Conserv 5:245–249 National Parks and Wildlife Service, Sydney Isaac NJB, Mallet J, Mace GM (2004) Taxonomic inflation: its Spencer RJ, Georges A, Welsh M (2007) The Bellinger Emy- influence on macroecology and conservation. Trends Ecol dura. Ecology, population status and management. Report Evol 19:464–469 to NSW National Parks and Wildlife Service, Sydney Iverson J (1992) A revised checklist with distribution maps of Spinks PQ, Thomson RC, Shaffer HB (2009) A reassessment of the turtles of the world. Earlham College, Richmond, IN Cuora cyclornata Blanck, McCord, and Le, 2006 (Tes- King P, Heatwole H (1994) Non-pulmonary respiratory sur- tudines: Geoemydidae) and a plea for taxonomic stability. faces of the chelid turtle Elseya latisternum. Herpetologica Zootaxa 2018:58–68 50:262–265 Spinks PQ, Thomson RC, Shaffer HB (2010) Nuclear gene Kuchling G, Rhodin AG, Ibarrondo BR, Trainor CR (2007) A phylogeography reveals the historical legacy of an ancient new subspecies of the snakeneck turtle Chelodina inland sea on lineages of the western pond turtle, Emys mccordi from Timor-Leste (East Timor) (Testudines: Cheli- marmorata in California. Mol Ecol 19:542–556 dae). Chelonian Conserv Biol 6:213–222 Swofford DL (1996) PAUP*: Phylogenetic analysis using par- Lauck T, Clark CW, Mangel M, Munro GR (1998) Implement- simony (*and other methods), Version 4. 0d59. Sinauer ing the precautionary principle in fisheries management Associates, Sunderland, MA through marine reserves. Ecol Appl 8:72–78 van Dijk P, Stuart B, Rhodin A (2000) Asian turtle trade. Pro- Legler J (1960) A simple and inexpensive device for trapping ceedings of a workshop on conservation and trade of aquatic turtles. Proc Utah Acad Sci 37:63–66 freshwater turtles and tortoises in Asia. Chelonian Res Lips KR, Brem F, Brenes R, Reeve JD and others (2006) Monogr 2:1–164 Emerging infectious disease and the loss of biodiversity in Villesen P (2007) FaBox: an online toolbox for FASTA se - a Neotropical amphibian community. Proc Natl Acad Sci quences. Mol Ecol Notes 7:965–968 USA 103:3165–3170 Walters CJ, Holling CS (1990) Large-scale management Mace GM (2004) The role of taxonomy in species conserva- experiments and learning by doing. Ecology 71: tion. Philos Trans R Soc Lond B Biol Sci 359:711–719 2060–2068 McCord W, Cann J, Joseph-Ouni M (2003) A taxonomic Walther GR, Post E, Convey P, Menzel A and others (2002) assessment of Emydura (Testudines: Chelidae) with de- Ecological responses to recent climate change. Nature scriptions of new subspecies from Queensland, Australia. 416:389–395 Reptilia 27:59–63 Wilson EO, Peter FM (eds) (1988) Biodiversity. National Acad- McGlothlin KL (2004) Loss of biodiversity. Rowman & Little- emy Press, Washington, DC field Publishers, Lanham, MD World Resources Institute (1992) Global biodiversity strategy: McKelvey KS, Aubry KB, Schwartz MK (2008) Using anecdo- guidelines for action to save, study, and use Earth’s biotic tal occurrence data for rare or elusive species: the illusion wealth sustainably and equitably. World Resources Insti- of reality and a call for evidentiary standards. Bioscience tute, World Conservation Union (IUCN), and United 58:549–555 Nations Environment Programme, Washington, DC Moritz C (1994) Defining evolutionarily significant units for Yamaura Y, Amano T, Kusumoto Y, Nagata H, Okabe K conservation. Trends Ecol Evol 9:373–375 (2011) Climate and topography drives macroscale biodi- NSW NPWS (New South Wales National Parks and Wildlife versity through land-use change in a human-dominated Service) (2001) Bellinger River Emydura (Emydura mac- world. Oikos 120:427–451 quarii) recovery plan. NSW NPWS, Sydney Zhu D, Degnan S, Moritz C (1998) Evolutionary distinctive- Oliver PM, Lee MSY (2010) The botanical and zoological ness and status of the endangered Lake Eacham rainbow- codes impede biodiversity research by discouraging pub- fish (Melanotaenia eachamensis). Conserv Biol 12:80–93

Editorial responsibility: Michael Mahony, Submitted: February 1, 2011; Accepted: April 7, 2011 Callaghan, NSW, Australia Proofs received from author(s): June 7, 2011